Re: Spinlocks and compiler/memory barriers

Robert Haas <robertmhaas(at)gmail(dot)com> writes:
> Despite my concerns about keeping the list of supported atomics short,
> and I do have concerns in that area, I'm not really sure that we have
> much choice but to go in that direction. We can't accept a >5x
> performance hit in the name of portability, and that's literally what
> we're talking about in some cases. I definitely want to think
> carefully about how we proceed in this area but doing nothing doesn't
> seem like an option.

To be clear, I'm not advocating doing nothing (and I don't think anyone
else is). It's obvious based on Andres' results that we want to use
atomics on platforms where they're well-supported. The argument is
around what we're going to do for other platforms.

regards, tom lane